HUGE DOMAINS SNIPING GODADDY CLOSEOUTS

[wwwweb]

So annoying Godaddy hasn't stopped Huge Domains from sniping Godaddy Closeouts with their automated tools, no way a human bidder can win a even closeout.

First they were sniping with the backorders, now you cut that out, and you are letting them snipe via automated tools.

So what do you say @Joe Styler , you want to even the playing field a bit, as your partners are bidding everything in a split second, from $12, to $11, and bidding everything else into the hundreds from a simple bid. I would rather pay a Huge Domains surcharge at checkout.


Huge Domains has an unfair advantage on the auction platform, essentially taxing every user for using it with their automated access advantages given to them thru the house.

 

[manpreet]

That's for registrars to sell Afternic names. I mean to automate listing my domains. AFAIK the closest option is the csv upload but that's still manual and because you have to reload your whole list to remove a name or change pricing it can take some time to process after upload. I'm writing my own script to keep track of all my names. Would be nice to have it use an api to automate listing my names, changing prices over time or removing names sold outside AN.

I think only Afternic staff has the control over that, let me know too if you find a solution otherwise.

 

[Randolph]

GD says they get the same API limits as anybody else, which is 60 requests per minute. Now, this is actually entirely possible. I could test this out and prove it by building a bot which would adhere to the same rules if I had an excess of time and resources.

It IS possible that HD are not cheating as far as API access goes, at least.

Possible, not entirely probable.

Assuming it takes 3-5 API calls to research and fully bid a domain — within the typical drop auction window of 3-5 hours — one account is able to touch about 2500 domains daily with typical API limits. This is being generous.

HugeDomains can't only be touching 2500 domains during peak to get their turnouts; Custom limits or multiple accounts are the only way to attain their results.

Secondly, they have offices next door to godaddy. It is somewhat widely considered that these two have some under the table deals, which I'm inclined to believe, given what I've observed in the past. But this might also mean that their servers are very close together, and based on how this actually works intrinsically, might give them a natural (not unfair) advantage over anyone else.

I was investigating this for a short time but ultimately gave up as it got complicated.

Server proximity will not give that large of an advantage today. I work with high I/O, multi-touch (database pulls, etc) API calls and can relay calls in less than 50-80ms — from servers on the opposite sides of the globe. Servers in the same building have trouble breaking 40ms;

Data is essentially instantaneous. Hardware is the limitation.

I've been investigating from the standpoint of creating a "people-powered" drop catch platform, where everyone in the network is a "single unit" working together against the big catchers. I've talked with [current and ex] employees of the big names and inadvertently found a number of the industry secrets.

The industry plays under the notion that we don't know how software/closed-door deals/etc works and "customers are stupid."

I notice 2 bidder id's, one ends in 32, the other 33

These are obviously owned by the same entity; Meaning we must be allowed to own multiple accounts to circumvent the limits.

 

[7363824]

These are obviously owned by the same entity; Meaning we must be allowed to own multiple accounts to circumvent the limits.

I haven't actually read the TOS () but even if it does include a prohibition against multiple accounts setting up a separate LLC for each account so it is a legally distinct entity isn't that hard or even expensive, unlike setting up multiple accredited registrars - which we all know they do.

 

[Randolph]

I haven't actually read the TOS () but even if it does include a prohibition against multiple accounts setting up a separate LLC for each account so it is a legally distinct entity isn't that hard or even expensive, unlike setting up multiple accredited registrars - which we all know they do.

If GoDaddy allows separate LLCs by the same entity to create accounts with the sole intent of funneling to the main entity, they must allow everyone to create multiple accounts or no one.

 

[Joe Styler]

Afternic API available if you become their partner, you may apply here:

https://www.afternic.com/premium-partner-levels

right you have to be a partner i.e. reseller for an Afternic api we do not provide them to customers.

 

[tonyk2000]

The industry plays under the notion that we don't know how software/closed-door deals/etc works and "customers are stupid."

Analysing big data is the key. Sometimes I sincerely wish some "domain leaks" papers to become public, similar to what happens or happened in politics ("Panama Papers").

I've talked with [current and ex] employees of the big names

"For nothing is secret that will not be revealed, nor anything hidden that will not be known and come to light." (the New King James Bible). Some domain industry players should not forget this quote...

 

[7363824]

If GoDaddy allows separate LLCs by the same entity to create accounts with the sole intent of funneling to the main entity, they must allow everyone to create multiple accounts or no one.

(Insert a disclaimer that I am not a lawyer here...): An LLC is legally deemed a "person" from a court's perspective and is a separate entity from any other LLC or it's owner.

 

[Lord Antares]

Not in case of closeouts... (?) unless I'm missing something. How to check what bidder ID (not whois) grabbed closeout?

I notice 2 bidder id's, one ends in 32, the other 33

As Tony stated, could be another API for closeouts, we simply don't know as the whois is all the same.

Yes, my apologies. I confused it with expired auctions.

Assuming it takes 3-5 API calls to research and fully bid a domain — within the typical drop auction window of 3-5 hours — one account is able to touch about 2500 domains daily with typical API limits. This is being generous.

HugeDomains can't only be touching 2500 domains during peak to get their turnouts; Custom limits or multiple accounts are the only way to attain their results.

I don't follow. Are the closeout times randomized after a domain auction ends? Because if they're not, it would be a cake walk to pick out two dozen and win them.

Possible, not entirely probable.


Server proximity will not give that large of an advantage today. I work with high I/O, multi-touch (database pulls, etc) API calls and can relay calls in less than 50-80ms — from servers on the opposite sides of the globe. Servers in the same building have trouble breaking 40ms;

How? 40 ms seems unusually high.

If GoDaddy allows separate LLCs by the same entity to create accounts with the sole intent of funneling to the main entity, they must allow everyone to create multiple accounts or no one.

You see, this is interesting.

In a personal message to me, Joe Styler said using multiple devices is a firm NO. However, several pages back in this thread, in reply to how HD are so successful in catching closeouts, he said ''maybe they are using multiple devices''. Take it as you may.

 

[Randolph]

(Insert a disclaimer that I am not a lawyer here...): An LLC is legally deemed a "person" from a court's perspective and is a separate entity from any other LLC or it's owner.

This is true to an extent. Which is why I included "with the sole intent of funneling to the main entity."

Legally, if you do your personal dirty work under your LLC, it all falls on you at a personal level as you are essentially the same entity. Same concept applies here otherwise the playing field is on a hill.

I would be 100% on-board if HugeDomains had a separate entity called MediumDomains and kept the domains completely separate.

 

[tonyk2000]

I don't follow. Are the closeout times randomized after a domain auction ends? Because if they're not, it would be a cake walk to pick out two dozen and win them.

Hard to say. It depends on... who knows what. Actually, GD is buggy (especially after the last upgrade) and one would reasonably guess that auctions part should also be affected from time to time.

 

[wwwweb]

I see a lot of people don't even bother fighting anymore in the auctions, it is basically like terminator, and skynet, the machines are running the place other than the few gems, or other top picks people are after. Huge Domains is spending a good portion of their incoming funds to keep buying domains, 6 million strong, and they keep growing.

 

[tonyk2000]

I would be 100% on-board if HugeDomains had a separate entity called MediumDomains and kept the domains completely separate.

And their business model may even allow one entity to bid against another, why not?

Of course we are speaking about conspiracy theories here. Nothing more. No accusations. It would be up to the individual reader to do their own research and decide whether or not it was valid.
*end of legal stuff*

 

[dirk]

How? 40 ms seems unusually high.

I agree. That seems absurdly high to me as well. And let's not forget, every single ms counts.

Proximity does definitely matter.

 

[Randolph]

How? 40 ms seems unusually high.

Sub-40 round trip is rarely seen for these types of requests under normal constraints.

Most definitely can get lower than these times when unencrypted, no outside API/DB, and/or infrastructure is properly tuned, which is rarely the case.

You see, this is interesting.

In a personal message to me, Joe Styler said using multiple devices is a firm NO. However, several pages back in this thread, in reply to how HD are so successful in catching closeouts, he said ''maybe they are using multiple devices''. Take it as you may.

This is one reason why I attempted to get an answer publicly. We should all see how level the playing field is since "[n]o one is doing anything anyone else cannot do."

Just put a large amount of cash into a reserve and drive auction price up hundreds of percents and you too can be shrugged at.

 

[Randolph]

I agree. That seems absurdly high to me as well. And let's not forget, every single ms counts.

Proximity does definitely matter.

I agree that proximity matters but "will not give that large of an advantage."

I only work with large sized, dynamic calls and full encryption due to the nature of the data.

When you work with highly redundant systems that always spit at an average delay, you can time calls ahead of time to fire and complete within a few ms of your target time.

 

[dirk]

I agree that proximity matters but "will not give that large of an advantage."

I only work with large sized, dynamic calls and full encryption due to the nature of the data.

When you work with highly redundant systems that always spit at an average delay, you can time calls ahead of time to fire and complete within a few ms of your target time.

True, but I would beg to differ that it doesn't give them a big advantage. Using common sense,there's a reason they're that close. That's no coincidence.

Secondly, you can time ahead all you want but there's no way knowing ahead what the RTT will be exactly. The smaller the distance, the more predictable/stable the RTT will be. That seems like a major advantage to me.

 

[Lox]

something like this?
... stable multiple (cpu) servers which can handle higher traffic and different communications methods beyond just request-response <=> routing & authentication server (reg. Api ID) <=> GD API server/s.

 

[7363824]

Good chance their servers are in data centers not their headquarters...I mean they could have build a data center in their headquarters but that's not usually how its done...just saying.

 

[Randolph]

True, but I would beg to differ that it doesn't give them a big advantage. Using common sense,there's a reason they're that close. That's no coincidence.

Secondly, you can time ahead all you want but there's no way knowing ahead what the RTT will be exactly. The smaller the distance, the more predictable/stable the RTT will be. That seems like a major advantage to me.

I agree, my main point is that no matter the latency, it is hard to pull the numbers of HD within the limits of the public API.

I just wanted everyone's suspicions confirmed because it's technically hearsay and conspiracy. Though, I think we all have an idea of what's going on in the back of our minds.


I'll say this because you'll probably be interested:

For the last few years, I've been working on predictive infrastructure with big data startups. It works almost like Google's Stadia.

We build and hold a ledger, cached on the edge, that generates predictions based on factors such as parse size, distance to servers, real-time connection speed, queue, CPU/RAM loads, etc. With this, we make inferences by comparing historic data to the current request.

We can predict with high accuracy what each dynamic call will entail, how long it will take, as well as predicting the typical flow of request-responses. ie: If Request-X can pull 200 variances of data in subsequent requests, but an overwhelming number of those requests pulled only 30 variances in the past, we can call these 30 in advance and have the data indexed or cached, ready to be served before a formal request was actually made.

One interesting startup took this and increased data center efficiency exponentially by ordering and timing requests to have a nominal effect on their infrastructure and active parses. They were able to do this because we can predict RTT down to single digit ms majority of the time.

(Typical AI, ML, neural net, buzzword here, buzzword here)

 

[Randolph]

something like this?
... stable multiple (cpu) servers which can handle higher traffic and different communications methods beyond just request-response <=> routing & authentication server (reg. Api ID) <=> GD API server/s.

Yeah, tough to achieve sub-40ms, even between neighboring server racks.